JPH03191673A - Black/white edge processing method for color reading device - Google Patents

Abstract

PURPOSE:To prevent the occurrence of the drag of a black/white edge and the false colors and to obtain a read picture of high quality by comparing the difference of outputs between the picture elements having an interval secured between them with the set value with use of the sensor output read by a transparent filter. CONSTITUTION:For the picture element signal obtained from the main scan of a line image sensor 5, the output differences X between the sensor output (n) obtained by a transparent filter 9-1 and the sensor output (n+2) of a picture element separated from the picture element signal with an interval secured between them are successively calculated. Then no edge is decided when the absolute value of the difference X is smaller than the threshold value S. While an edge is decided vice versa. When the edge deciding result is obtained, the white and black levels are defined to the larger and smaller ones respectively between two sensor outputs obtained by the Y and R filters 9-2 and 9-3 set at the same positions as the picture element (n) and (n+1) used for calculation of the difference X. Furthermore the two sensor outputs obtained by both filters 9-2 and 9-3 set at the same positions as an intermediate picture element (n+1) between picture elements (n) and (n+2) are replaced with the white or black levels in accordance with the plus or minus code of the difference X.

Description

[Detailed Description of the Invention] [Summary] This invention relates to a black-and-white edge processing method for a color reading device that reads color images by switching multiple filters in a time-sharing manner, which prevents drooping of black-and-white edges and the occurrence of false colors, thereby producing high-quality read images. In the main scanning direction, the difference between every other pixel output is determined for the W filter output, and if the absolute value of the difference is greater than or equal to a predetermined value, it is determined to be an edge, and the larger one is white, and the smaller one is the other side is black;
Furthermore, the intermediate position is replaced with black or white depending on the sign of the difference. On the other hand, in the sub-scanning direction, the output difference of every other line pixel in the sub-scanning direction is calculated for the W filter output for three lines at the same pixel position, and if the absolute value of the difference is equal to or greater than a predetermined value, it is judged as an edge. The larger one is set to white, the smaller one is set to black, and intermediate positions are replaced by black or white depending on the sign of the difference.

Furthermore, the outputs of the Y and R filters are made to match the processing results of the W filter.

[Industrial Application Field] The present invention relates to a black-and-white edge processing method for a color reading device that reads color addition by switching a plurality of filters in a time-division manner.

Color reading devices are used as image input devices for facsimiles and computers, and are used to input images such as R (red), G (green),
A device that reads three colors of B (blue) using a CCD image sensor or the like is known. These color reading devices are desired to be further downsized, lowered in price, and to have higher precision in read images, and in particular, to be able to reproduce black and white edges clearly.

[Prior Art] FIG. 10 shows the basic reading structure of a conventional filter switching type color reading device.

In FIG. 10, a reading line 6 on a document 1 is illuminated by an illumination device, such as a fluorescent lamp 2, and an image on this line 6 is read. The reflected light from the reading line 6 is imaged onto a CCD line image sensor (hereinafter referred to as rCCD line sensor) 5 by a reduction optical system including a lens 4. Such a configuration is a general black and white reading optical system.

In order to perform color reading, between the lens 4 and the sensor 5, three colors for color separation, for example, R, G, and 83 R filters 9a, G filters 9b, and B filters 9c are provided. The R, G, and B filters 9a, 9b, and 9c are held by linear motors 10a, 10b, and 10c, respectively, and can be moved to positions where they block the optical path from the reading line 6 to the sensor 5.

1) Figures (a) to (c) show color reading operations. When reading the R (red) component of the color image 6 on the reading line, drive the linear motor 10a of the three filters to insert the R filter 9a into the optical path as shown in Figure 1) (a). do. The other two filters 9b and 9c are held at positions where they do not block the optical path. As a result, of the reflected light from the reading line 6, only the red light can reach the CCD line sensor 5, and the R component of the image is read.

When reading the G (green) and B (blue) components of an image, a G filter 9b and a B filter 9c are inserted into the optical path, respectively, as shown in FIG. 1) (b) and (c). When one line of RGB reading is completed, the document is sent a predetermined length in the direction of the arrow in FIG. 10, and the next line is read.

By combining the small and lightweight filters 9a to 9c with the high-speed linear motors 10a to 10c, the filters can be switched in approximately the same amount of time as the reading time of one sensor line (1 to 10m5). Achieve color reading.

Although the filter 9 is inserted between the lens 4 and the CCD light image sensor 5 in Section 10.1), the same effect can be obtained even if the filter 9 is inserted between the document 1 and the lens 4.

Figure 12 shows another color reading structure. In this color reading structure, three color R, G, and B filters are arranged as a single plate so as to move up and down in the optical path. ing. A specific filter switching structure is shown in FIG. 13, in which R, G, and B filters 9a, 9b, and 9c are fixed to a base plate 21. The base plate 21 is a plate spring 2
It is held on a support stand 24 via 2. Coils 23 are attached to both sides of the base plate 21, and a magnet (not shown) is arranged on the fixed side of the coil 23. When the coil 23 is energized, the base plate 21 is driven in the vertical direction, and the filters 9a to 9 are
C can be switched into the optical path plane.

In order to perform color reading with the switching structure shown in FIG. 12.13, it is necessary to move the R, G, and B filters 9a, 9b, and 9c back and forth once in the optical path. First, R filter 9a
reads the R signal so that the G filter 9b enters the optical path, and then moves the base plate 21 so that the G filter 9b enters the optical path.
Read the G signal. Furthermore, move the B filter 9c in the same manner.
Complete signal reading and one line reading. Thereafter, the base plate 21 is moved so that the R filter 9a enters the optical path again, and the R signal is read, and this is repeated thereafter.

[Means for Solving the Problems] In such a conventional color reading device, a fluorescent lamp and a CCD line sensor are used for boat fishing.

This is due to the relatively low price and the ability to read lines at high speed. However, in the color manuscript,
When there is a black and white image, for example, when reading characters, line drawings, graphs, etc., there is a problem in that colors that should not be present appear at the edges of the image. Hereinafter, this will be explained separately in the main scanning direction, which is the reading scanning direction of the CCD line sensor, and the sub-scanning direction, which is the document feeding direction.

As an example, a transparent filter (W), a yellow filter (Y), and a red filter (R), which constitute complementary color filters, are used as filters.

The CCD line sensor has, for example, 2,048 14 μm square sensors (light receiving pixels) lined up. When a pixel of such a CCD line sensor straddles the edge of a monochrome image during reading, the output of that pixel is neither white nor black, but is an intermediate level output.

In addition, line illumination using fluorescent lights has a wider illumination range than the spot illumination of a drum scanner, so interference from light reflected from the document surface around the reading position may cause
There is also the problem that the edges of the image become blurred. In addition, blurring occurs due to limits in lens resolution, resulting in blurred image edges.

FIG. 14 shows the CCD output when reading the edges of an image.

That is, when the light-receiving pixel of the CCD reads an edge as shown in FIG. 14(a), a CCD output as shown in FIG. 14(b) is obtained.

In FIG. 14(b), the nth pixel is at the black level, and the n+th pixel is at the black level.
The level of the pixel increases due to the blur, the n+2 pixel outputs an intermediate level at the black and white edge, the n+3 pixel decreases the level due to the blur, and the n+4 pixel outputs a white level. According to such a CCD sensor output, the actual read image becomes an image with sagging edges in the main scanning direction. Furthermore, if the appearance of these droop changes for each of the three colors of color separation due to chromatic aberration of the lens, a color that should not exist (false color) occurs in the drooped edge area.

Next, the droop in the sub-scanning direction, which is the document feeding direction, will be explained.

First, in the filter switching color reading device shown in FIG. 10.12, three colors are read in a time-division manner. Normally, the document is continuously fed in the sub-scanning direction.

FIG. 15 shows the state of reading in the sub-scanning direction, and the relative positions of the original and the sensor change due to movement of the original or movement of the reading illumination system or sensor.

FIG. 15(a) shows a state in which the W filter is reading a black and white edge as case 1, and the COD output by each color filter at this time. Also, Fig. 15(b), (
c) shows cases 2 and 3 in which the Y filter or R filter is reading black and white edges, and the CCD output of each color filter.

Even in the sub-scanning direction, blurring due to lighting occurs in the same way as in the main-scanning direction, so in all cases 1 to 3, the black level may be slightly different between black (mth line) and gray (m+2th line). There is an increased output (line m+1). According to the output of such a sensor, in the actual read image, colors that should not exist (false colors) occur at the black and white edges in the sub-scanning direction.

As described above, when reading the edges of a monochrome image by switching the filter, as shown in Figure 16, edges appear in both the main scanning and sub-scanning directions, or colors that should not exist appear in the reading. There was a problem that the image deteriorated.

The present invention has been made in view of these conventional problems, and therefore, it is possible to prevent blurring of black and white edges and the occurrence of false colors in the main and sub-scanning directions of color scanning when filters are switched, thereby achieving high quality. It is an object of the present invention to provide a black and white edge processing method for a color reading device that allows a read image to be obtained.

[Means to solve the problem] FIG. 1 is a diagram explaining the principle of the present invention.

First, as shown in FIG. 1Cfa), the present invention focuses the reflected light from the original 1 on the line image sensor 5 through a reading optical system such as a mirror or lens, and synchronizes it with the reading of one line. The object of the present invention is a color reading device that reads a color image by time-divisionally switching a plurality of filters provided in the optical path of the reading optical system.

In the black and white edge processing method of the present invention for such a color reading device, the plurality of filters include a transparent filter (W filter) 9-L that transmits all wavelengths of visible light and two other color separation filters 9-L. 2.9-3, for example, a yellow filter (Y filter) that transmits yellow light and a red filter (R filter) that transmits red light are provided, and the pixel signal obtained by main scanning of the line image sensor 5 is Processing is performed as shown in Figure 1 (b).

In the processing in the main scanning direction in FIG. 1(b), (1) Regarding the sensor output n from the transparent filter 9-1, calculate the output difference ΔX between the sensor output (n + 2) of the pixel that is one pixel apart. Sequential calculations are performed, ■When the absolute value of the output difference ΔX is smaller than a preset threshold S, it is determined that it is not an edge and no edge processing is performed, and when it is larger than the threshold S, it is determined that it is an edge. When the determination result is obtained, regarding the two sensor outputs from the two pixels n and n+2 based on the calculation of the output difference ΔX and the other two filters 9-2 and 9-3 at the same position as these two pixels, The larger one is the white level, and the smaller one is the black level. (2) Further, the intermediate pixel n+1 of the two pixels n and n+2 used to calculate the output difference ΔX, and the two filters 9-2 and 9-3 located at the same position as the intermediate pixel. The two sensor outputs are replaced with a white level or a black level depending on the positive or negative sign of the output difference ΔX.

In addition, in the present invention, as shown in FIG. 1(c), as a processing method in the sub-scanning direction which is the document feeding direction, (1) Regarding the sensor output m of each pixel by the transparent filter 9-1, The difference ΔY from the sensor output m+1 of the line one line apart is calculated sequentially, and if the output difference ΔY is smaller than a preset threshold T, it is determined that it is not an edge, and no edge processing is performed. When it is larger than , it is judged as edge, and when the judgment result of edge is obtained,
2 pixels m, m+2 of 2 lines used for calculation for output difference ΔY
, and the two sensor outputs from the other two filters 9-2.9-3 located on the same line and at the same position as the two pixels,
The larger one is the white level, and the smaller one is the black level. ■Furthermore, the pixel m+1 on the middle line of the two lines used to calculate the output difference ΔY, and the other two funuil filters on the same line and at the same position as the middle line pixel. 2 according to 9-2.9-3
The two sensor outputs are replaced with a white level or a black level depending on the sign of the output difference ΔY.

Here, when replacing the intermediate pixel n+1 in the main scanning direction or the intermediate line pixel m+1 in the sub-scanning direction, for example, when the output difference ΔX, ΔY is positive, it is replaced with a white level, and when it is negative, it is replaced with a black level.

[Function] According to the black-and-white edge processing method for a color reading device according to the present invention having such a configuration, complementary color filters, such as transparent (white) filter, red filter, and yellow filter, and in the main scanning direction, use the sensor output read by the transparent filter, compare the difference in output of pixels with one interval with the set value, and calculate the size relationship. If it is determined to be an edge, as opposed to a black and white edge, each of the three colors of read data (9 pixels in total) for the compared pixel and the pixels sandwiched between the compared pixels is set to white or black. This removes the edge color.

In addition, in the sub-scanning direction, regarding the sensor output read by a transparent filter, the difference between the output of a certain pixel and the pixel at the same position in one line is compared with the set value, and it is determined whether it is a black and white edge based on the magnitude relationship. However, if it is determined to be an edge, each of the three colors of read data (total of 9 pixels) is set to white or black for a total of three lines including the compared line and the lines sandwiched between the compared lines.

By this processing, droop and false colors occurring at the edges of the read image can be removed to obtain a clear image.

[Embodiment] FIG. 2 is a diagram showing the configuration of a color reading device to which the black and white edge processing method of the present invention is applied.

In FIG. 2, 30 is a reading optical system, for example, a third reading optical system.
It has the color reading configuration shown in the figure.

The color reading configuration diagram in FIG. 3 is similar to the conventional example in FIG.
The reading line 6 on the original 1 is illuminated by an illumination device, such as a fluorescent lamp 2, and the reflected light from the reading line 6 is imaged on the CCD line sensor 5 by the lens 4.
It is read as a line pixel signal. Lens 4 and CCD
Three color separation filters 9-1 are installed between the line sensors 5.
．． 9-2.9-3 are arranged, and in the present invention, these 3
One of the color separation filters is a transparent (white) filter that transmits all wavelengths of visible light (hereinafter referred to as "W filter").
) 9-1, and the other two filters 9-2
．． The filters 9-3 are a complementary color yellow filter (hereinafter referred to as "Y filter") 9-2 and a red filter (hereinafter referred to as "R filter") 9-3. Of course, W, Y,
Each of the R filters 9-1 to 9-3 is switched and inserted into the reading optical path in a time-division manner by near motors 10a, 10b, and 10c, so that pixel signals obtained through three filters are obtained for one line of reading. ing.

Referring again to FIG. 2, the image of the document reading line obtained from the reading optical system 30 is captured by a CCD provided in the reading circuit 32.
The line sensor 5 converts the reading signals into a number corresponding to the number of pixels in one line, and the A/D conversion circuit 34 converts the signals into digital data, which is then converted into a 3-line memory 4 capable of storing 3 lines.
0 to the processing circuit 50. In the processing circuit 50, a main scanning direction processing unit 36 processes black and white edges in the main scanning direction, which is the reading scanning direction of the CCD line sensor 5, and a sub-scanning direction processing unit 38 processes black and white edges, which is the document feeding direction. Processing of black and white edges in the sub-scanning direction is performed to create highly accurate pixel signals from which blur and false colors at black and white edges are removed.

The read signal processed by the processing circuit 50 regarding black and white edges is converted into complementary colors w and y by the color conversion circuit 42.

The R signal is converted into three color signals of R, G, and B and output.

In the embodiment shown in FIG. 2, color reading is performed by switching filters as shown in FIG. 3, but one CCD
Regarding light source switching type color reading using the line sensor 5 and three color light sources, the black and white edge processing method of the present invention can be applied as is by including a white light source corresponding to the W filter 9-1 among the three color light sources. Can be applied. Furthermore, it is effective not only in color reading but also in monochrome reading to remove out-of-focus images and blurring of image edges caused by illumination.

Next, a black and white edge processing method by the main scanning direction processing section 36 shown in the processing circuit 50 of FIG. 2 will be described with reference to the operational flow diagram of FIG. 4.

Regarding the processing in the main scanning direction shown in FIG.
Assume that the CCD output from the W filter 9-1 shown in a) is obtained. When the CCD output from the W filter 9-1 shown in FIG. 5(a) is subjected to the processing in the main scanning direction shown in FIG. 4, the processing result shown in FIG. 5(b) is obtained.

In FIG. 4, first step SL (hereinafter referred to as "step")
W for one line to be processed in (omitted),
CCD outputs from each of the R and Y filters are read from the 3-line memory 40.

Next, the process proceeds to 82, and the output difference ΔX between the n-th CCD output currently being processed and the n+2-th CCD output one space apart is calculated for the CCD output of the W filter.

Next, the process proceeds to 83, where the absolute value of the output difference ΔX calculated in S2 is compared with a preset threshold value S. When the absolute value of the output difference ΔX is smaller than the threshold S, edge processing is not performed and S1
If the value is 0, the process proceeds to the next (n+1)th pixel. On the other hand, when the absolute value of the output difference ΔX is larger than the threshold value S in S3, the process proceeds to S4 and it is determined that there is an edge.

In the case of FIG. 5(a), the absolute value of the output difference ΔX between the n-th pixel and the n+2-th pixel is smaller than the threshold S, so in this case, the process proceeds to 810 and the next n+1-th pixel is Move to processing. For the next n+1 pixel, n=n+1 in S2, so the n+1 pixel and 1
Calculate the difference between the n+3rd CCD outputs with an interval of
3, it is determined whether the absolute value of the output difference ΔX is larger than the threshold value S. In this case, if it is established that the absolute value of the output difference is greater than the threshold value S, the process proceeds to S4 and it is determined that there is an edge. If an edge determination result is obtained in S4, n+
The larger one of the first and n+3rd CCD outputs is set as the white level, and the smaller one is set as the black level. Next, the process proceeds to S5, where the n+1st and n+3rd Y filters 9-2 and R
The CCD output from the filter 9-3 is set to the same level as the white level and black level of the CCD output from the W filter 9-1 at the same position set in S4.

Next, proceed to 86 and proceed to n+1 which was not used in the calculation of S2.
In order to determine the n+2-th black and white level between the n+3-th and n+3-th black and white levels, it is determined whether the sign of the output difference ΔX calculated in S2 is + or -. If the sign is 10, the process proceeds to S7, where the intermediate (n+2)th W filter 9-1 and Y filter 9-2 are processed.
and each CCD output of the R filter 9-3 are all replaced with white level. On the other hand, if the sign is - in S6, S
8, each CCD output of the intermediate (n+2)th W filter 9-1, Y filter 9-2, and R filter 9-3 is replaced with a black level. In the case of FIG. 5(a), since the sign of the output difference ΔX between the n+1st and n+3rd in S2 is -, the process proceeds to 88, where the intermediate n+2nd W, Y,
It is assumed that each CCD output of the R filters 9-1 to 9-3 has a black level.

As a result, the CCD with the W filter shown in Fig. 5(a)
As shown in the same figure (b), the output is from the n+2 pixel to the n+
The reading signal can be made to change from the black level to the white level at the third pixel without any intermediate stage, and other Y filters 9
Since the CCD outputs of the -2 and R filters 9-3 are also the same as those of the W filter 9-1, it is possible to reliably prevent drooping and false colors from occurring at the edge portions.

When the processing for black and white edges for one line is completed through the processing from 81 to S8, it is determined at 89 that the processing for one line has ended, the next line is set at S1), and the processing from 81 to S8 is repeated again.

Next, a black and white edge processing method in the sub-scanning direction, which is the document reading direction, will be described with reference to the operation flowchart of FIG.

In FIG. 6, first, 3 line memory 40 to 3
Assume that the CCD outputs of the W, Y, and R filters for a line are read and, for example, data shown at Dl in FIG. 7 is obtained.

This read data D1 is data when a black and white edge is read by the W filter 9-1 shown in FIG. , the CCD output of the R filter is obtained. Note that black in parentheses is an output whose level has increased somewhat from the black level, and white in parentheses is an output whose level has decreased somewhat from the white level.

Next, proceed to S2, regarding the CCD output of the W filter,
Calculate the output difference ΔY between the m-th line and the m+2-th line.

Next, the process proceeds to 83, where it is determined whether the absolute value of the output difference ΔY is greater than a preset threshold T. If it is smaller than the threshold T, the process advances to SIO and the next pixel is processed. On the other hand, S3
If the absolute value of the output difference ΔY is larger than the threshold T, that is, if the condition shown in D2 in FIG. The larger one of the CCD outputs is defined as a white level, and the smaller one is defined as a black level. Next, proceed to 85, line m and m
CC by Y and R filters at the same position on the +2nd line
The D output is replaced with the same white level or black level as the CCD output of the W filter determined in S4.

Next, the process proceeds to S6, in order to determine the black and white level of the intermediate line m+1, the sign of the output difference ΔY calculated in S2 is + or -.
Determine whether If the sign is 0, proceed to S7, and the intermediate m
The CCD outputs of the W, Y, and R filters on the +1st line are all set to white level.

If it is determined in S6 that the sign is -, the process proceeds to S8, and the CCD outputs of the W, Y, and R filters on the m+1th intermediate line are replaced with the black level.

In the case of FIG. 6, if the sign of the output difference ΔY between the m-th line and the m+2 line is, for example, 10, then the intermediate m+
The first line is replaced with a white level in S7.

With the above processing, C in the sub-scanning direction shown as Dl in FIG.
As for the CD output, as shown in D3, the m-th line is all black level, the m+1 and m+2 lines are all white level,
It is possible to remove false color and droop. When the processing in S7 or S8 is completed, the process advances to 89, where it is determined whether or not the processing for one line has been completed. If the processing for one line has not been completed, the process returns to SIO and proceeds to the processing of the next pixel. When the processing for the line is completed, the process advances to Sll to update to the next line, and the same processing is repeated thereafter.

FIG. 8 shows, as case 2, the processing details in the sub-scanning direction of FIG. 6 when black and white edges are read with a Y filter. In other words, when a black and white edge is read with a Y filter, the 15th
The case is as shown in Figure (b), and the W filter 9 in this case
The CCD output of −1 is as shown in Dl, and the condition that the absolute value of the output difference between the mth and m+2nd is larger than the threshold T is satisfied in D2, and as a result, the mth and m+1th are both black level, For the m+2th image, a processing result in which all the images are at the white level is obtained.

Moreover, FIG. 9 shows the case where black and white edges are read with the R filter, that is, case 3 shown in FIG. 15(c).
When black and white edges are read with R filter 9-3, D
CCD output of m−m+3 lines shown in l is obtained, and D2
Then, a condition is established that the absolute value of the difference between the CCD outputs of the m+1st line and the m+3rd line is larger than the threshold T,
As a result, as shown in D3, a processing result is obtained in which the m, m+l, and m+2 lines are all black, and the m+3 line is all white.

This type of processing in the main scanning direction and sub-scanning direction prevents intermediate levels from occurring in the edges and edges of monochrome images, thereby preventing the occurrence of false colors due to image blurring and chromatic aberration, and achieving high-precision reading. .

Here, the threshold value S1 for edge determination in the main scanning direction and the threshold value T for edge determination in the sub-scanning direction vary depending on the optical system, circuit system, etc., but the optimal values can be determined while actually reading edges. is desirable.

[Effects of the Invention] As explained above, according to the present invention, when a black and white image is read by a color reading device, it is possible to reliably eliminate blurring and false colors that occur in both the main scanning direction and the sub-scanning direction. The quality of the read image is guaranteed, and highly accurate reading can be achieved.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram of the principle of the present invention; FIG. 2 is a diagram of the device configuration of the present invention; FIG. 3 is a color reading configuration diagram of the present invention; FIG. 4 is a flowchart of the operation in the main scanning direction of the present invention; 6 is an explanatory diagram of processing in the main scanning direction of the present invention; FIG. 6 is a flowchart of the operation in the sub-scanning direction of the present invention; FIG. 8.9 is an explanatory diagram of the sub-scanning direction processing of the present invention; FIG. 10 is a conventional color reading configuration diagram;
32: Reading circuit 1) Figure is an explanatory diagram of filter switching operation in Figure 10; 34: A/D converter Figure 12 is a conventional color reading configuration diagram;
36. Main scanning direction processing section FIG. 13 is a block diagram of the switching mechanism shown in FIG. 12; 38. Sub-scanning direction processing section 14
The figure is an explanatory diagram of reading black and white edges in the main scanning direction; 40
:3 line memory FIG. 15 is an explanatory diagram of reading black and white edges in the sub-scanning direction; 42: Color conversion circuit FIG. 16 is an explanatory diagram of droop that occurs on black and white edges.50. Processing circuit. In the figure, 1: Original 2: Fluorescent lamp (light source) 4: Lens 5: CCD line image sensor 6: Reading line 9-1 = Transparent filter (W filter) 9-2: Yellow filter (Y filter) 9-3 = Red filter (R filter) 10a to 10C: Linear motor 30: Reading optical system

Claims (3)

[Claims] (1) The reflected light from the original (1) is imaged on the line image sensor (5) via a reading optical system such as a mirror or lens, and the optical path of the reading optical system is synchronized with the reading of one line. In a color reading device that reads a color image by time-divisionally switching a plurality of filters provided therein, the plurality of filters include a transparent filter (9-1) that transmits all wavelengths of visible light, and two other color separations. Filter (9
-2, 9-3) is provided, and for each pixel signal obtained by main scanning of the line image sensor (5), the sensor output (n) of the transparent filter (9-1) is determined by The output difference (ΔX) with the sensor output (n+2) is calculated sequentially, and when the absolute value of the output difference (ΔX) is smaller than the preset threshold (S), it is determined that it is not an edge and no edge processing is performed. , when it is larger than the threshold value (S), it is determined as an edge, and when the edge determination result is obtained, the output difference (S) is determined as an edge.
The two pixels (n, n+1) used in the calculation of Δx) and the other two filters (9-2, 9-3) at the same position as the two pixels
Regarding the two sensor outputs, the larger one is the white level,
The smaller one is the black level; Furthermore, the two pixels (n, n
+2) regarding the two sensor outputs from the intermediate pixel (n+1) and the two filters (9-2, 9-3) located at the same position as the intermediate pixel, depending on the sign of the output difference (ΔX). A black and white edge processing method for a color reading device, characterized in that the white level or the black level is replaced. (2) The reflected light from the original (1) is imaged on the line image sensor (5) via a reading optical system such as a mirror or lens, and the optical path of the reading optical system is synchronized with the reading of one line. In a color reading device that reads a color image by time-divisionally switching a plurality of filters provided therein, the plurality of filters include a transparent filter (9-1) that transmits all wavelengths of visible light, and two other color separation filters. Filter (9
-2, 9-3), and the transparent filter (9-1
), the difference (ΔY) between the sensor output (m) of each pixel at the same pixel position and the sensor output (12) of one line apart is calculated, and the absolute value of the output difference (ΔY) is When it is smaller than a preset threshold (T), it is determined that it is not an edge and no edge processing is performed, and when it is larger than the threshold (T), it is determined that it is an edge, and when the edge determination result is obtained. is the output difference (Δ
2 pixels (m, m+2) of 2 lines used for calculation of Y),
Regarding the sensor outputs of the two pixels from the other two filters (9-2, 9-3) located on the same line and at the same position as the two line pixels, the larger one is the white level, the smaller one is the black level, and furthermore, the above-mentioned The pixel (m+1) on the middle line of the two lines used to calculate the output difference (ΔY) and the other two filters (9-2, 9-
2) A black and white edge processing method for a color reading device, characterized in that the sensor outputs of the two pixels are replaced with a white level or a black level depending on the sign of the output difference (ΔY). (3) The intermediate pixel (n+1) in the main scanning direction or the pixel (m+1) on the intermediate line in the sub-scanning direction is replaced with a white level when the output difference (Δx) or Δ(Y) is positive. 3. A black and white edge processing method for a color reading device according to claim 1, wherein when , is negative, it is replaced with a black level.